Net Reclassification Improvement Analysis
| Model with LVEF, log BNP, and H/M | |||
| Model with LVEF and log BNP | Low probability (<8%) | Intermediate probability (8%–24%) | High probability (>24%) |
| Subjects with events | |||
| Low probability (<8%) | 9 | 4 | 0 |
| Intermediate probability (8%–24%) | 2 | 50 | 9 |
| High probability (>24%) | 0 | 3 | 50 |
| Subjects without events | |||
| Low probability (<8%) | 180 | 29 | 0 |
| Intermediate probability (8%–24%) | 71 | 347 | 29 |
| High probability (>24%) | 0 | 28 | 113 |
Net gain in reclassification (event) = 8/127 = 0.063 (P = 0.059). Net gain in reclassification (no event) = 41/797 = 0.051 (P = 0.001). Net reclassification improvement: 0.114 (P = 0.002). Table includes only subjects with results for all 3 variables.